1. Field of the Invention
This invention relates to a method and an apparatus for embedding or inserting a watermark into contents data such as picture data, audio data, audio visual data, or multimedia data. In addition, this invention relates to a method and an apparatus for reproducing or extracting a watermark from watermarked contents data.
2. Description of the Related Art
H. Ogawa et al. have reported “A Copying Information Embedding Method using DCT for Digital Movies”, SCIS'97-31G, which discloses watermarking methods suitable for MPEG bit streams. The watermarking methods are based on modifying DCT coefficients, motion vectors, and quantizer matrices. Regarding contents data into which watermark information have been inserted by the DCT-coefficient-modifying watermarking method reported by H. Ogawa et al., the accuracy of reproducing the watermark information from the contents data is considerably low when the synchronization with DCT-sampling units (DCT blocks) fails to be acquired. The failure of the acquisition of the synchronization causes DCT blocks to shift from true positions by, for example, one pixel or one line in a frame. It is assumed that picture data which have been watermarked by one of the methods reported by H. Ogawa et al. are subjected to affine transformation to slightly rotate or distort a picture. It tends to be difficult to accurately extract the watermark from the transformation-resultant watermarked picture data.
J. Ohnishi et al. have reported “A Watermarking Scheme to Image Data by PN Sequence”, SCIS'97-26B, which discloses a data hiding method using a PN sequence in the spread spectrum technique. In the data hiding method, an original image signal is converted into a spread spectrum in response to a PN sequence. A narrow band signal to stand for a signature is added to the spread spectrum, that is, a wideband channel of which an original image is spread. When the signature-added spread spectrum is inversely converted into the normal image by the PN sequence, the signature signal is spread over the normal-image signal. In other words, the signature signal is embedded in the normal-image signal. The spread signature signal is low in power, and hence hardly acts as noise with respect to the original image. Accordingly, the signature-added image is substantially the same as the original image. When the signature-added image signal is spread by the PN sequence, the signature signal is reproduced.